Catal Lett (2008) 122:1-8
DOI 10.1007/s10562-007-9382-4
An Investigation of the Reduction and Reoxidation of Isolated
Vanadate Sites Supported on MCM-48
Jason L. Bronkema Æ Alexis T. Bell
Received: 25 October 2007 / Accepted: 3 December 2007 / Published online: 20 December 2007
Ó Springer Science+Business Media, LLC 2007
Abstract The reduction and subsequent reoxidation of
isolated vanadate species supported on silica was investigated using temperature-programmed reduction and
oxidation, along with in-situ XANES and Raman spectroscopy. Approximately 70–80% of the vanadium was
reduced to V3+ after reduction in H2 at temperatures up to
923 K. Upon reduction, the vanadyl oxygen was removed
and the three remaining V–O bonds are lengthened by 0.2
Å. The vanadate species are rapidly reoxidized when
exposed to O2, with the amount of oxygen uptake matching
well with the amount removed during reduction. In-situ
Raman spectroscopy during reoxidation in 18O2 showed
that significant scrambling occurs between gas phase
oxygen and surface oxygen species during the reoxidation
of the vanadate species.
Keywords
Vanadate Silica Reduction Oxidation
1 Introduction
Supported vanadium oxide catalysts have been shown to be
active for the partial oxidation of methane and methanol to
formaldehyde, the oxidative dehydrogenation of light
alkanes, and the selective reduction of NOx. Interest in
understanding the mechanism and kinetics of these reactions has motivated studies of the structure of the vanadate
J. L. Bronkema A. T. Bell
Chemical Sciences Division, Lawrence Berkeley National
Laboratory, Berkeley, USA
J. L. Bronkema A. T. Bell (&)
Department of Chemical Engineering, University of California,
Berkeley, CA 94720-1462, USA
e-mail: [email protected]; [email protected]
species and of the dynamics of reduction and oxidation of
the supported vanadia. The structure of vanadia on the
surface of metal oxide supports such as silica, alumina,
titania, and zirconia has been studied using infrared
and Raman spectroscopy [1–3], UV-Visible spectroscopy
[3, 4], 51V NMR [5], XANES, and EXAFS [6]. Application
of these techniques has shown that depending on the surface
density of vanadium, vanadia can be present as isolated
monovanadate species, two-dimensional polyvanadate
species, and bulk V2O5. At very low surface densities (less
than about 0.2 V/nm2), vanadium is present exclusively as
isolated vanadate species. As the surface density of vanadium increases polyvanadate species are formed. Finally, as
monolayer coverage by polyvanadate species is approached, nanoparticles of V2O5 are observed.
Several investigators have examined the reduction of
supported vanadate species. Kanervo et al. [7]. have
reported that samples containing mainly isolated monovanadate species could be reduced by 50% from V5+ to
V3+; however, at higher vanadium loadings, they obtained
reductions of 62%. A broad TPR peak was observed in the
range of 600–800 K, with the maximum for samples containing exclusively monovanadate species occurring at
750 K. Work by Koranne et al. on V/SiO2 and V/Al2O3
showed that 70–90% of the V5+ could be reduced to V3+ in
5% H2 in Ar by raising the temperature up to 1,173 K [8].
Similar levels of reduction have also been reported by Nag
and Massoth [9]. Finally, Argyle et al. have shown using
transient-response UV-visible spectroscopy that under
conditions of propane oxidative dehydrogenation the rate
of vanadia reduction for alumina-supported vanadia is 104
slower than the rate at which the partially reduced vanadia
is reoxidized [10]. It is noted, though, that in none of these
studies was the structure and oxidation state characterized
before and after reduction and after reoxidation.
123
2
Consequently, understanding the dynamics of the reduction
and oxidation of well-defined vanadate species remains a
subject of interest.
The objective of this investigation was to obtain a fundamental understanding of the dynamics of reduction and
reoxidation of isolated monovanadate species supported on
silica. Silica was chosen as the support because it is relatively inert. The loading of vanadium was kept low so that
all of the vanadium would be present as isolated monovanadate species. XANES and EXAFS were used to
determine the oxidation state of vanadium and the local
geometry of the vanadate species before and after reduction, and following reoxidation. The dynamics of vanadia
reduction and oxidation were investigated using temperature-programmed reduction and oxidation (TPR and TPO).
Additional information about the structure of vanadia
during reduction and reoxidation was obtained by in-situ
Raman spectroscopy.
2 Experimental Methods
A high-surface area mesoporous silica, MCM-48, was
prepared following procedures reported in the literature
[11, 12]. Vanadium oxide was deposited onto MCM-48 by
chemical vapor deposition of vanadium acetylacetonate.
The weight loading, as determined by elemental analysis,
was 3.43 wt.% vanadium and the surface area of the catalyst was 1,370 m2/g as determined by analysis of the BET
isotherm. The apparent surface density of vanadium was
0.3 V/nm2.
Raman spectra were recorded using a Kaiser Optical
HoloLab series 5,000 Raman spectrometer equipped with a
Nd:YAG laser that is frequency-doubled to 532 nm. The
laser was operated at a power level of 25 mW measured at
the sample with a power meter. A Princeton Instruments
TEA/CCD detector was operated at -40 °C. Spectra were
recorded with a resolution of 2 cm-1. Samples (approximately 25 mg each) were pressed into pellets at 5,000 psi
and placed within on a rotating sample holder in a quartz
Raman cell. The samples were rotated at 100 rpm during the
measurements to reduce the effects of sample heating by the
laser. During reoxidation, the samples were heated in
30 cm3 min-1 flow of 5% O2/He. The temperature was
increased from 323 to 773 K at 4 K/min. Raman spectra
were collected every 75 s in order to obtain a scan every 5 K.
X-ray absorption (XAS) measurements were performed
at the Stanford Synchrotron Radiation Laboratory (SSRL)
on beam lines 2–3. The measurements were performed at
the vanadium K-edge. The edge energy for each sample
was determined as the first inflection point of the main peak
in the spectrum, and the edge energy of the vanadium foil
was set to 5,465 eV. The optimal sample amount was
123
J. L. Bronkema, A. T. Bell
calculated based upon the weight fraction of all atomic
species to obtain an absorbance of *2.5, which corresponds to 14 mg of catalyst, with 30–40 mg of boron
nitride added to make a stable pellet. Samples were placed
in a controlled-atmosphere cell that allowed for heating up
to 823 K in the presence of flowing gas [13]. UHP He and
10% O2 in He were used in these experiments to pretreat
the catalyst samples before the XAS measurement. After
the pretreatment was completed, the cell was evacuated to
6 x 10-4 Pa and cooled to 77 K before the XAS data were
collected. For in-situ XANES measurements, scans were
made approximately every 7 min at reaction temperature
and under gas flow.
The EXAFS portion of the XAS data was analyzed to
determine the local coordination around each vanadium
atom. The software program IFEFFIT, along with its
complementary GUIs Athena and Artemis, were used for
the data analysis [14, 15]. First, a linear pre-edge was
subtracted from the data, fit to the range of -150 to
-75 eV relative to the edge energy. Then a quadratic
polynomial fit to the range of 150–850 eV for EXAFS and
100–300 eV for in-situ XANES, both relative to the edge
energy, was used to normalize the data. A cubic spline was
fit to the normalized data from 2 to 16 Å-1 to remove the
background absorption and minimize the signal in R space
below 1 Å. These data were then Fourier transformed over
the k range of 2.7–9.5 using the Kaiser-Bessel window
functions with a dk of 1 and a k weight of 3. The EXAFS
fits were performed in R space, with data fit from 1.0 to 3.0
Å. The value for So2, determined by performing a fit on
vanadium foil, used in the EXAFS analysis was 0.7.
Temperature-programmed reduction and oxidation were
carried out using 50 mg of catalyst placed in a quartz
microreactor heated by a small furnace. TPR was carried
out by flowing 2% H2 in He at 60 cm3/min while heating the
reactor from 298 to 923 K at 10 K/min and monitoring the
effluent by mass spectrometry. TPO was carried out in a
similar manner using 0.5% O2 in He flowing at 60 cm3/min.
The effluent from the reactor was analyzed using a MKS
Cirrus quadrupole mass spectrometer.
3 Results and Discussion
A detailed analysis of the local structure of the V/MCM-48
catalyst in the oxidized state has been published previously
[6]. The catalyst consists of isolated vanadate species in a
pseudotetrahedral geometry and all of the vanadium is in
the 5+ oxidation state. EXAFS analysis shows that the
sample contains one V=O bond with a length of 1.59 Å and
three V–O bonds with a length of 1.80 Å.
Figure 1a shows the TPR of the V/MCM-48 sample
upon heating in 2% H2/He to 923 K at 10 K/min. No H2 is
Reduction and Reoxidation of Isolated Vanadate Sites Supported on MCM-48
1.00
3
1.2
A
1.0
Normalized Intensity
µmol/s
0.75
0.50
0.25
0.8
0.6
0.4
0.2
300
400
500
600
700
800
900
Temperature (K)
0.0
5460
5480
5500
5520
5540
Energy (eV)
0.12
B
Fig. 2 In-situ XANES scans during TPR of V/MCM-48. Sample was
heated at 10 K/min to 823 K and held for 1 h under a 10% H2/He
flow. The arrows show the direction of changes in the spectrum as the
reduction temperature increases
0.10
µmol/s
0.08
0.06
0.04
0.02
0.00
300
400
500
600
700
800
900
Temperature (K)
Fig. 1 (a) TPR of V/MCM-48 catalyst in 2% H2/He at 10 K/min. (b)
Comparison of hydrogen uptake and water desorption during TPR.
Black-hydrogen, red-water
consumed when a similar experiment was performed with
MCM-48 alone. Reduction starts at 600 K and reaches a
maximum rate of reduction at 860 K. Integrating the
amount of hydrogen consumed and normalizing it to the
amount of vanadium in the sample gives a ratio of H2
consumed per V of 0.85. This suggests that most but not all
of the vanadium centers are reduced from V5+ to V3+. It is
also observed in Fig. 1b that the hydrogen consumption
and water desorption peaks are roughly equivalent, suggesting that each H2 removes one oxygen atom from the
vanadium to produce one H2O.
The TPR spectrum presented in Fig. 1 consists of two
distinct peaks. Deconvolution of these features was performed to determine their positions and relative intensities.
The first peak is broad, is centered at 810 K, and comprises
74.5% of the total area. The second peak is much sharper,
is centered at 850 K, and comprises 25.5% of the total area.
The positions of these peaks are similar to those reported
by Koranne et al., who suggested they are due to the
reduction of oligomeric vanadia [8]. This interpretation is
inconsistent, though, with the results of Raman spectroscopy, which shows no traces of V2O5, and EXAFS analysis
of the oxidized catalyst, which detects no evidence for V-V
backscattering as would be expected if significant surface
concentrations of oligomers were present. Hence, a more
likely interpretation is that the two peaks seen in Fig. 1 are
attributable to the reduction of distorted and undistorted
pseudo-tetrahedrally coordinated monovanadate species.
Figure 2 shows XANES spectra of V-MCM-48 while it
is heated at 10 K/min to 823 K in 10% H2/He and then
held at this temperature for 1 h. The position of the edge
energy, which is defined as the maximum of the first
derivative of the main edge, contains information about the
oxidation state of the sample. Prior to the onset of reduction, the edge energy of the V/MCM-48 is 5,484.0 eV,
indicating that all of the vanadium is in the 5+ state. As
reduction proceeds, the edge energy shifts to lower energies, with the largest changes in the edge energy occurring
over the temperature range of 650–823 K.
The XANES data presented in Fig. 2 were analyzed
with a linear combination fitting algorithm in the Athena
software in order to determine the extent of reduction
during each scan [15]. The standard for the V5+ species was
the first scan in the experiment, since the catalyst is fully
oxidized under these conditions, and bulk V2O3 was used
as the standard for V3+. The results of the fitting are plotted
in Fig. 3a. It is observed that the sample consists of almost
entirely V5+ species up to 680 K. Above this temperature,
the sample reduces rapidly. The results of this experiment
show that after the sample has been reduced in H2 at 823 K
for 1 h, 85% of the V5+ cations have been converted to V3+
cations. This value is higher than 0.71 H2/V that was calculated from a TPR experiment carried out with a heating
123
4
J. L. Bronkema, A. T. Bell
A 100
Oxidation state (%)
80
60
40
20
0
300
400
500
600
700
800
Temperature (K)
B
5.0
Average Oxidation State
4.8
4.6
4.4
4.2
4.0
3.8
3.6
3.4
3.2
3.0
300
400
500
600
700
800
900
1000
Temperature (K)
Fig. 3 (a) Linear combination fitting results of XANES during TPR
in 10% H2/He at 60 cm3/min. (b) Oxidation State versus Temperature
during TPR experiments. Black-TPR data in Fig. 1, red-XANES TPR
data in Fig. 2
between experiments, 2% H2 during TPR and 10% H2
during XANES. However, the more likely cause is that
while the heating rates in both experiments are the same,
the temperature measurement in the TPR reactor is more
accurate than that in the cell used for XAS measurements.
As a result, the latter data are believed to provide a more
accurate representation of the change in the average oxidation state of vanadium with reduction temperature.
EXAFS data, obtained for V/MCM-48 after it had been
reduced under the conditions used to acquire the XANES
data, are shown in Fig. 4. The Fourier transform of k3v(k)
was fitted using the Artemis software. In carrying out this
analysis it was assumed that the reduced vanadate species
is bonded to silica by three equivalent V–O–si bonds [16].
The best fit to the data shows that the sample contains three
V–O bonds of length 2.02 Å. This suggests that upon
reduction, the V=O oxygen is removed and that the V–O
bonds lengthen from 1.79 Å to 2.02 Å. The increase in
length of the V–O bond is consistent with what is seen
upon comparison of the V–O bond lengths in V2O5 and
V2O3. V2O5 consists of one V=O bond of length 1.58 Å
and four V–O bonds with lengths between 1.78 and 2.02 Å.
By contrast, V2O3 has three V–O bonds that are 2.02 Å in
length and three V–O bonds that are 2.12 Å in length. The
Debye-Waller factor (r2) increases to 0.0085 upon reduction, suggesting that the reduced sites are more disordered
than the fully oxidized sites, for which r2 was only 0.0014.
The peak centered at 2.6 Å in the plot corresponds well to a
V-Si backscattering path of length 3.40 Å. No evidence
was seen for V-V backscattering in the reduced sample,
indicating that the vanadium centers remain isolated upon
reduction. The R-factor for the fit of the reduced sample is
0.08. The lack of a better fit is very likely due to the
5.0
123
2.5
3
FT of k χ(k)
protocol identical to that used for the acquisition of
XANES data. The reason for the slight increase in the
calculated amount of reduction calculated by the linear
combination fit is likely due to small differences in the
XANES of reduced V/MCM-48 compared to V2O3. Small
changes in the edge or pre-edge features of reduced
V/MCM-48 would lead to relative changes in the extent of
reduction. However, since the sample was never fully
reduced, the best standard available for use is bulk V2O3.
Plots of the average oxidation state of V/MCM-48 as a
function of reduction temperature determined from the
TPR spectrum presented in Fig. 1 and the analysis of the
XANES data presented in Fig. 2 are shown in Fig. 3b.
While the two methods for determining the average oxidation state are similar, the curve based on the XANES
data exhibits a shaper decrease in oxidation state with
temperature than the curve based on the TPR data. This
could be due to differences in the hydrogen concentration
0.0
-2.5
-5.0
0
1
2
3
4
5
R (Å)
Fig. 4 Experimental and fitted EXAFS data for reduced V/MCM-48.
Black—magnitude of experimental data, red—magnitude of fit,
green—imaginary part of experimental data, blue—imaginary part
of fit
Reduction and Reoxidation of Isolated Vanadate Sites Supported on MCM-48
5
1.2
0.020
1.0
Normalized Intensity
µmol O2 /s
0.015
0.010
0.8
0.6
0.4
0.005
0.2
0.000
300
0.0
400
500
600
700
800
5460
5480
100
5540
A
80
Oxidation state (%)
assumption that the reduced sample contains only V3+
species, whereas the actual extent of reduction for the
sample is 0.71–0.85. Thus, the experimentally observed
EXAFS pattern contains some oxidized centers as well as
reduced centers.
Figure 5 shows the TPO spectrum of the V/MCM-48
after it had been reduced using the protocol described
above for the XANES experiments. The oxidation of the
reduced sample begins at 300 K and reaches completion
near 650 K. The total amount of oxygen consumed was 0.7
O/V, which suggests that the sample was about 70%
reduced to V3+ at the beginning of the experiment. This
value agrees very well with the calculated hydrogen consumption (0.71 H2/V) during conditions identical to the
initial reduction of the sample.
The total amount of oxygen needed to reoxidize the
catalyst was also determined by pulsed oxygen chemisorption. For these experiments, the sample was reduced
under conditions identical to the XANES TPR experiments. After purging in helium, pulses of oxygen were sent
to the sample to determine the amount of oxygen necessary
to completely reoxidize the catalyst. The results show that
between 0.72 and 0.76 O/V are necessary to fully reoxidize
the catalyst. This number is in excellent agreement with the
amount of oxygen uptake determined by TPO (see above).
XANES data were acquired during reoxidation of the
catalyst. Figure 6 shows the normalized XANES data
during reoxidation, while Fig. 7a shows the results of the
linear combination fitting of these spectra. The standards
for V3+ and V5+ were the same as those used to represent
the XANES data obtained during reduction. It is noted that
a small amount of reoxidation is likely to have occurred in
the 10 min during which the EXAFS cell was flushed with
O2 at 293 K. Therefore, at the start of the temperature
ramp, the sample is only 60% reduced. Reoxidation occurs
5520
Fig. 6 In-situ XANES during TPO of V/MCM-48 in a flow of 10%
O2/He at 60 cc/min. The arrows show the direction of the changes in
the spectrum as the oxidation temperature increases
60
40
20
0
300
400
500
600
700
800
Temperature (K)
5.0
B
4.8
Average Oxidation State
Fig. 5 TPO of V/MCM-48 in 0.5% O2/He at 10 K/min. Sample was
initially reduced in H2 at 823 K for 1 h
5500
Energy (eV)
Temperature (K)
4.6
4.4
4.2
4.0
3.8
3.6
3.4
300
400
500
600
700
800
Temperature (K)
Fig. 7 (a) Linear combination fitting results of XANES spectra
during TPO. (b) Oxidation State versus Temperature during TPO
experiments. Black-TPO data in Fig. 5, red-XANES data in Fig. 6
123
6
123
A
υ V=O
2000
5
υ V=O, V O
800
400
Increasing Temp.
1200
2
υSi-O
Intensity (arb. units)
1600
0
300
600
900
1200
Wavenumber (cm-1)
3500
B
3000
Intensity (arb. units)
rapidly and is completed between 650 and 700 K, in very
good agreement with the results of the TPO described
above. It was also observed that the XANES spectrum of
the sample at the end of the experiment matches that for the
fully oxidized sample recorded prior to reduction. This
observation demonstrates that the sample was fully
reoxidized.
Figure 7b shows the average oxidation as a function of
the reoxidation temperature determined from the TPO data
of Fig. 5 and from the XANES data of Fig. 6. While
qualitatively similar, the average oxidation state determined from the XANES data is higher than that determined
from the TPO data. As in the case of Fig. 3b, the differences are likely to differences in the accuracy of the
temperature measurements and in the flow rates of O2
relative to the weight of the catalyst used in the two types
of experiment.
In-situ Raman observations were made in order to
observe the changes occurring in the supported vanadate
species during the reduction and the subsequent reoxidation of the catalyst. Shown in Fig. 8a are backgroundsubtracted Raman spectra taken during the reoxidation of
V/MCM-48 previously reduced in the Raman cell for 1 h at
773 K in flowing H2. Due to changes in the sample color,
and therefore the sampling depth of the laser, the intensity
of all bands decreased during reduction. Since silica does
not reduce under the conditions used to reduce the supported VOx species, the silica feature at 500 cm-1 was used
to normalize the data at each temperature, as shown in
Fig. 8b. It is observed in this figure that the intensity of the
V=O band at 1,033 cm-1 is lower after the reduction,
suggesting that O atoms in V=O bonds are consumed upon
reduction. During reoxidation, this feature is restored to its
original height prior to reduction. Above 670 K, a small
band is observed at 996 cm-1, which is likely due to the
formation of a small amount of V2O5 upon reoxidation.
The formation of small nanoparticles of V2O5 can be
ascribed to the hydration of VOx species in the presence of
water produced during reduction and oligomerization of
these species during reoxidation. This interpretation is
suggested by hydration/dehydration experiments reported
by Xie et al. [17]. To determine the amount of V2O5
formed, the two peaks in the spectrum were deconvoluted
assuming Lorenzian peak shapes, and the V2O5 area was
corrected for the ten-fold higher Raman scattering crosssection compared to that of isolated VO4 species [17]. By
this means it is estimated that only 0.5% of the vanadium
formed V2O5.
The same experiment was performed using 18-labeled
oxygen in an attempt to gain further insights into the process by which gas-phase oxygen reoxidizes the isolated
vanadium centers. Figures 9a and 9b show that in the
presence of 18O2, the principal feature reappearing in the
J. L. Bronkema, A. T. Bell
2500
2000
1500
1000
500
0
300
600
900
1200
-1
Wavenumber (cm )
Fig. 8 Raman spectra during reoxidation in 16O2 after reduction in
H2 for 1 h at 773 K. (a) Background subtracted spectra. (b) Spectra
normalized by SiO2 band at 500 cm-1. Scans in increasing temperature: black-310 K; red-370 K; green-430 K; blue-490 K; light blue550 K; pink-610 K; gold-670 K; dark blue-730 K; purple-773 K
Raman spectrum is the band at 1,033 cm-1 for V=16O and
only a small band is seen at 987 cm-1 for V=18O. These
spectra were deconvoluted in order to determine the relative fraction of V=18O/V=16O, assuming that the Raman
scattering cross-sections for both species were identical,
and the results are presented in Table 1. The fraction 18Olabeled V=O groups is initially 2.5% and increases by the
end of the experiment to 14.8%. The total area of the V=O
band returns to the original area, indicating that the sample
is completely reoxidized. These observations suggest that
16
O in the silica support can replace the O atoms removed
from V=O species in the isolated vanadate structures.
A similar observation has been reported earlier by Ohler
and Bell, when isolated molybdate species are first reduced
in H2 and then reoxidized in 18O2 [18].
A mechanism by which reduced vanadate centers could
be oxidized by O2 has been proposed recently by Goodrow
and Bell as a part of their theoretical study of methanol
Reduction and Reoxidation of Isolated Vanadate Sites Supported on MCM-48
2000
reduced center, resulting in the formation of a vanadium
peroxo species. This species then undergoes rearrangement
so that one of the two O atoms forms a V=O bond, whereas
the second O atom forms a V–OO–si structure. In the next
step the peroxo group moves to form a Si–OO–si species,
similar to that observed to form during the migration of
oxygen in silica [19, 20]. Migration of the peroxo group
continues until it encounters a second reduced vanadate
species, at which point the peroxo species is used to oxidize the vanadium in this group back to V5+ and form a
second V=O bond. If 18O2 is used for reoxidation, it is easy
to understand how the small amount of 18O2 required to
reoxidize the sample can scramble with the significantly
larger pool of 16O present on the surface of silica.
Assuming that 18O and 16O scramble completely, it is
estimated that 4.1% of the V=O bonds upon reoxidation
should contain an 18O label. This figure is in reasonable
agreement with what is observed experimentally.
1500
4 Conclusions
1200
16
O
A
υ V=
18
Increasing Temp.
υ V=
800
O
υSi-O
Intensity (arb. units)
1000
600
400
200
0
300
600
900
1200
-1
Wavenumber (cm )
2500
Intensity (arb. units)
7
B
1000
500
0
300
600
900
1200
-1
Wavenumber (cm )
Fig. 9 Raman spectra during reoxidation in 18O2 after reduction in
H2 for 1 h at 773 K (a) Background subtracted spectra (b) Spectra
normalized by SiO2 band at 500 cm-1. Scans in increasing temperature:/time black-473 K; red-573 K; green-673 K; blue-773 K; light
blue-2.5 min; pink-5 min; gold-7.5 min; dark blue-10 min; purple12.5 min
oxidation to formaldehyde over isolated vanadate species
supported on silica [16]. The reoxidation scheme that they
propose begins with the rapid adsorption of O2 by the
Table 1 Relative percentage of 18O in V=O bonds determined by
deconvolution of Raman spectra shown in Fig. 9
Temp/Time
Percent
473
0.0
573
2.5
673
4.6
773
9.2
2.5 min
9.3
5 min
10.7
7.5 min
10 min
12.8
12.1
12.5 min
14.7
18
O (%)
The results of this study indicate that V5+ in isolated vanadate species supported on silica can be reduced by
approximately 70–80% to V3+ in flowing H2 at temperatures up to 923 K. XANES data confirm that the oxidation
state of vanadium in changes from V5+ to V3+ upon
reduction, and Raman spectroscopy indicates that the
oxygen in the V=O bond of the original vanadate group is
removed during reduction. Analysis of EXAFS data show
that the three V–O bonds that connecting the vanadate
species to the support lengthen by 0.2 Å after reduction.
Complete reoxidation of the reduced vanadate species can
be achieved with O2, and occurs at significantly lower
temperatures than does reduction. Raman spectroscopy
reveals that upon reoxidation of reduced vanadate species
using 18O2, considerable scrambling occurs between 16O
atoms in the support and those in the oxidizing gas.
Acknowledgments This work was supported by the Director, Office
of Energy Research, Office of Basic Energy Sciences, Chemical
Science Division, of the U.S. Department of Energy under Contract
No. DE-AC02-05CH11231. Portions of this research were carried out
at the Stanford Synchrotron Radiation Laboratory, a national user
facility operated by Stanford University on behalf of the U.S.
Department of Energy, Office of Basic Energy Sciences.
References
1. Deo G, Wachs IE (1994) J Catal 146:323
2. Wachs IE, Jehng JM, Deo G, Wechkuysen BM, Guliants VV,
Benzinger JB, Sundaresan SJ (1997) J Catal 170:75
3. Olthof B, Khodakov J, Bell AT, Iglesia E (2000) J Phys Chem B
104:1516
4. Burcham LJ, Deo G, Gao X, Wachs IE (2000) Top Catal
11/12:85
123
8
5. Wang C-B, Deo G, Wachs IE (1998) J Catal 178:640
6. Bronkema JL, Bell AT (2007) J Phys Chem C 111:420
7. Kanervo JM, Harlin ME, Krause AOI, Bañares MA (2003) Catal
Today 78:171
8. Koranne MM, Goodwin JG, Marcelin G (1994) J Catal 148:369
9. Nag NK, Massoth FE (1990) J Catal 124:127
10. Argyle MD, Chen K, Resini C, Krebs C, Bell AT, Iglesia E
(2004) J Phys Chem B 108:2345
11. Morey MS, Davison A, Stucky GD (1998) J Por Mat 5:195
12. Huo Q, Margolese DE, Stucky GD (1996) Chem Mater 8:1147
13. Jentoft RE, Deutsch SE, Gates BC (1996) Rev Sci Instrum
67:2111
14. Newville M (2001) J Synchrotron Radiat 8:322
15. Ravel B, Newville M (2005) J Synchrotron Radiat 12:537
16. Goodrow A, Bell AT (2007) J Phys Chem C 111:14753
123
J. L. Bronkema, A. T. Bell
17. Xie S, Iglesia E, Bell AT (2000) Langmuir 16:7162
18. Ohler N, Bell AT (2005) J Phys Chem B 109:23419
19. Skuja L, Güttler B (1996) Phys Rev Lett 77:2093; Orellana W, da
Silva AJR, Fazzio A (2001) Phys Rev Lett 87:155901; Bongiorno
A, Pasquarello A (2002) Phys Rev Lett 88:125901; Hoshino T,
Hata M, Neya S, Nishioka Y, Watanabe T, Tatsumura K,
Ohdomari I (2003) Jpn J Appl Phys Part 1 42:6535; Tatsumura K,
Shimura T, Mishima E, Kawamura K, Yamasaki D, Yamamoto
H, Watanabe T, Umeno M, Ohdomari I (2005) Phys Rev B
72:045205
20. Hamann DR (1998) Phys Rev Lett 81:3447; Ng KO, Vanderbilt
D (1999) Phys Rev B 59:10132; Hoshino T, Hata M, Nishioka Y,
Watanabe T, Tatsumura K, Ohdamari I (2003) Jpn J Appl Phys
Part 1 42:3560